Upload
ddro
View
242
Download
3
Tags:
Embed Size (px)
DESCRIPTION
OfDMA Calibration
Citation preview
STG(09)04
ECCElectronic Communications Committee
CE
PT
STG #17, WGSE - SEAMCAT Technical GroupERO, Copenhagen3-4 March 2009
Date Issued: 2 March 2009
Source : ERO
Subject: Part2 - Calibration test of the SEAMCAT DL OFDMA module
Document: for discussion
Password protection required? (Y/N) N
1 Background
In the last STG meeting #16 (27 November 2008), document STG(08)29rev1 presented the part 1 process for the calibration of the OFDMA module.
Throughout the part1, several errors where found in the beta version and a new 3.2.0 Beta8 containing the current development of the OFDMA module was released on 19 December 2008 and is accessible through http://www.seamcat.org/betaversion/. This was advertised on the SEAMCAT and STG mailing list.
2 Introduction
The calibration is based on the input from Qualcomm (STG(08)13 and STG(08)11). As a resuts of Part 2 work, SEAMCAT scenarios have been created to validate the implementation of the new OFDMA 3GPP LTE module:
Scenario 7For a first stage calibration, it is assumed that a traditional system (i.e. non-CDMA, non-OFDMA) will be employed as interferer with variable Tx power to consider the OFDMA self interference effect. This simulation are based on the assumptions and parameters of E-UTRA 10MHz macro cell system used in the Qualcomm simulator [1] and [2]. The impact from
1
interferer on victim system has been quantified in terms of throughput loss of E-UTRA versus Adjacent Channel Interference Power Ratio (ACIR) for DL.
The list of parameters for the SEAMCAT simulation are listed in Annex 4 Table 1 for DL case.
From the second run of calibration/testing, some discrepency have been identified. STG is invited to consider the results presented in this contributions and to take the aproppriate action.
Annex 1: input workspace scenario: STG(09)04 Annex1-scenario7.sws
Annex 2: UrbanMacro_OFDMA_calibration.class
Annex 3: Determination of the BS to BS (a.k.a. intersite) distance in SEAMCAT.
Annex 4: Simulation Assumptions and Parameters for DL
Annex 5: Propagation model plug-in description
Annex 6: Example of unwanted emission mask
3 Results of Simulation
3.1 Path loss
pathloss = 128.1 + 37.6 log10(2.5 km) + 10 *random ()[dB];
Propagation path loss (121.8+37.6*LOG10(d)+10)
0
20
40
60
80
100
120
140
160
0,1 0,15 0,25 0,35 0,45 0,55 0,65 0,75 0,85
distance d (m)
pat
h l
oss
(d
B)
Part 1 Part 2 - corrected propagation model
Figure 1: Propagation path loss based on the external evaluation using the original equation.
Path lossPart 2
Effective Path loss
Part 2
Figure 2:Effective Path loss, active links
Percentile level (%) Effective Path Loss for active links (dB)
0 7050 9099 113
Table 1: Summary of the effective path loss results for 0%tile, 50%tile and 99%tile
3.2 Received powerPart 2
(PBS=46 dBm)
Figure 3: SEAMCAT Results for the Rx power (using Tx max power = 46 dBm)
Rx power ranges from -64dBm to -24 dBm which is 46 dB than the path loss (70 dB to 110 dB) which is equivalent to the 46 dBm BS Tx power.
However, in the specification (STG(08)11), The Tx power from BS to the active UE is supposed to be divided by K the number of active UE:
Figure 4: Extract from STG(08)11 to set the transmit power assoicated to each UE
Action Point required: This need to be checked in the current implementation.
NOTE: In order to compensate for this inbalance, the BS maximum transmission power is set to 32 dBm for the continuation of the testing procedure.
Figure 5: Extract from STG(08)11 to calculate the received power in the serving cell/BS
Figure 6: SEAMCAT Results for the Rx power (C(j,k) in STG(08)11)
3.3 Inter System interference (Self interference)
Tx Power at BS [dB] Version 50% Self Interference [dB]46 Part 1 -8346 Part2 -7532 Part2 -88
Table 2: Summary results of the self interference from Part , Part 2 testing and for different input of the OFDMA BS Tx power.
Figure 7: Extract from STG(08)11 to calculate the self interference (Iother)
Part 2(with assumed PUE
BS=32 dB )
Figure 8: SEAMCAT Results for the Inter System Interference, active users(Iother(j,k) in STG(08)11)
Action Point required: Iother should be a function of the Tx power (arround 33 dB) and the path loss (varrying from ). Therefore the CDF should range with the value in Table 3. This need to be checked in the current implementation. Percentile level
(%)Effective Path Loss for active
links (dB)
Self Interference
0 70 32 -38
50 90 32 -5899 113 32 -81
Table 3: Summary of the effective path loss results for 0%tile, 50%tile and 99%tile
3.4 External interference
Figure 9: Extract from STG(08)11 to calculate the external interference (Iinter)
2.5 km => mean path loss = 143 dB (i.e. with 10dB log normal shadowing)Tx = 33 dBm, Tx gain = 0dBi, ACLR = 1000 dB
It follows the iRSS blocking value (dependent of the ACLR value entered)
Figure 10: External Interference, active users (ACLR = 1000 dB)
3.5 OFDMA Results
3.5.1 Introduction to ACIR in DL
Out-of-band-emissions maximum power: in the 3GPP UMTS – FDD mode the adjacent channel leakage ratio (ACLR) is defined by ACLRBS=PTX(in band)/PTX(out-of-band).
Maximum of adjacent channel power: the adjacent channel selectivity (ACS) is defined byACSUE=PRX(in band)/PRX(out-of-band).
In order to measure the combined effects of transmission and reception in the adjacent band, the DL adjacent carrier-to interference ratio ACIR is defined as:
Note: For then
then
In DL, the dominant part of ACIR is due to the UE frequency selectivity (ACS).
Figure 11: Illustration of frequency separation between victim and interferer
Therefore by varying the ACS of the victim receiver, it is possible to compare the results from STG(08)13 provided by Qualcomm.
Figure 12: Illustration of the selection of the ACS
The ACLR (at the interferer) is set to 1000 dB as shown below
Figure 13: Illustration of the selection of the ACLR (at the BS interferer)
3.5.2 SEAMCAT Results
This section (Table 4 and Table 5) will be filled in when self-interference implementation will be checked and approved by STG.
ACS 20 25 30 35 40Non int. Bitrate, MinimumNon int. Bitrate, MaximumNon int. Bitrate, AverageNon int. Bitrate, Std. Dev
Int. Bitrate, MinimumInt. Bitrate, MaximumInt. Bitrate, AverageInt. Bitrate, Std. Dev
Table 4: Summary of the capacity loss for variable ACS
ACIR [dB] Qualcomm SEAMCAT
throughput Loss in %
without shadowing correlation
throughput Loss in %
without shadowing correlation
20 6.9025 3.3030 1.4035 0.4040 0.20
Table 5: Throughput loss of downlink E-UTRA 10MHz system due to external interference for various ACIRs (ACLR = -1000 dB).
4 General remarks for developer
Elevation angle is always the same value, irrespective of the position of the UE to the BS.
The percentage should be rounded (see figure below)
Description missing for the SINR Minimum
Need description for the vectors in the red square (see figure below)
5 References
[1] STG(08)11 Benchmark for SEAMCAT OFDM Evolution: Simulation Assumptions and Methodology, Qualcomm.
[2] STG(08)13 Benchmark for SEAMCAT OFDM Evolution:Coexistence scenario: E-UTRA 10MHz to E-UTRA 10MHz Downlink
[3] STG(08)29rev1 part 1 - Calibration test of the SEAMCAT DL OFDMA module
Annex 3Determination of the BS to BS (a.k.a. intersite) distance in SEAMCAT.
There exists two method to defined hexagonal grid as shown in figure A1.1 below.
Qualcomm based method SEAMCAT based method
Figure A3.1: example of the two method to define an Hexagonal Grid and its results in SEAMCAT
Annex 4
Simulation Assumptions and Parameters for DL
Basic assumption: Macro cell environment with urban characteristics and adjacent channels.
1. Characteristic of the victim
The assumptions and parameters of E-UTRA 10MHz macro cell system used in the Qualcomm simulator and SEAMCAT are listed in table A4.1 for DL case. NC (No Change) in the SEAMCAT column means that the input value to SEAMCAT is the same as for the Qulacomm input.
Figure A4.1: General Victim tab to select the OFDMA module
Figure A4.2: General OFDMA tab to select general settings
1
Network Topology Qualcomm assumptions SEAMCAT equivalentNetwork Topology Hexagonal Grid, 19 sites (57
cells) with wrap around.NC – figure A4.6
Inter-Side Distance 750 m (i.e. Cell Range 500 m)
Cell Radius = 433 m – figure A4.6
Sectorization Tri- sector antennas defined in Annex 3
NC – figure A4.6
Base Station (BS) Antenna Gain
15 dBi NC- selection in figure A4.6 and definition in figure A4.7
User Equipment (UE) Antenna Gain
0 dBi NC – figure A4.6
Carrier Frequency 2000 MHz NC – figure A4.1Propagation Model Urban macro (BS-UE,
2000MHz)121.8+37.6*log10(d) dB (d is distance from BS to UE in kilometers)
Propagation plug-in used – figure A4.8
Log-normal Shadowing
10 dB without correlation 10dB to be added directly into the plugin as input param 1. This value is added to the randomization part of the lognormal shadowing– figure A4.8n
Minimum Coupling Loss
70 dB NC - figure A4.2
Bandwidth 10 MHz NC - figure A4.2Number of Available Resource Blocks (M)
24 NC - Max subcarriers per BS - figure A4.2
Number of Resource Block per UE (N)
1 NC - Number of subcarrier per mobile - figure A4.2
Number of Active UEs per Cell (K)
24 (K=M/N) NC - figure A4.5
Bandwidth of Resource Block
375 kHz (180kHz is the size of resource block later defined, but this has no big impact on the simulation results)
NC - figure A4.2
HO Margin 3 dB NC - figure A4.2Thermal Noise Density -174dBm/Hz Hard coded value in SEAMCAT = -
173.977UE Noise Figure 9 dB NC - figure A4.2Maximum BS Transmit Power
46 dBm NC - figure A4.4
Link Performance Model
Attenuated and truncated form of the Shannon bound
NC – Default value- figure A4.2
Table A4.1: Simulation assumptions and parameters of E-UTRA 10MHz DL as victim.
Figure A4.3: General OFDMA tab to select general settings
Figure A4.4: Link Specific tab
Figure A4.5: Capacity tab
Minimum subcarrier usage per Base Station:
When the system is filling the users in, the users would be added until the value exceed the minimum number of subcarrier usage per BS is reached (this value varies from the minimum to 100%). The default value is set to 100% for the calibration. STG decided that it is not useful to have such input parameter since for coexistence analysis the worth case scenario is to be considered and therefore a full loaded system should be used.
Figure A4.6: Positioning tab
Figure A4.7: Antenna Pattern at the Base Station
Figure A4.8: Propagation model tab
Figure A4.9: Path loss correlation tab
2 Characteristic of the Interferer
For a first stage calibration, it is assumed that a traditional system (i.e. non-CDMA, non-OFDMA) will be employed with variable Tx power to consider the OFDMA self interference effect.
Figure A4.10: Setting up the interfering frequency
Tx power 33 dBm (default value) scenario 7 (STG(09)04 Annex 1 - scenario 7.sws) with 1 snapshot
Figure A4.11: Setting up the Tx power of the interferer and the unwanted spectrum mask
Figure A4.12: coverage radius of the interferer (TX and Rx interferer)
Interfering link propagation model -> plug in
Figure A4.13: Using the propagation plug-in for the interfering link
Location of interferer to the victim
Figure A4.14: Setting up the position of the interferer (TXinterferer ) to the victim (Rxvictim ) (i.e. Simulation radius)
Figure A4.15: illustration of the position of the interferer to the victim using the traditional interface.
Figure A4.16: illustration of the position of the interferer to the victim using the OFDMA interface.
Annex 5
Propagation model plug-in
import java.util.Random;
import org.seamcat.model.propagation.PluginModel;
public class UrbanMacro_OFDMA_calibration implements PluginModel {
private Random random = new Random();
@Overridepublic double evaluate(double frequency, double distance, double TxHeight,
double RxHeight, int generalEnv, double logNormalShadowing, double param2, double param3) {
double pathloss = 128.1 + 37.6 * Math.log10(distance) + logNormalShadowing*random.nextGaussian();
return pathloss;}
}
Annex 6
Example of unwanted emission mask
Figure A6.1: Setting up the unwanted emission mask and extract the ACLR.